enhanced component performance study: emergency diesel

The INL is a U.S. Department of Energy National Laboratory operated by Battelle Energy Alliance

INL/EXT-19-54609

Enhanced Component Performance Study: Emergency Diesel Generators 1998–2018

Zhegang Ma

September 2019

NOTICE

This information was prepared as an account of work sponsored by an agency of the U.S. Government. Neither the U.S. Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for any third party’s use, or the results of such use, of any information, apparatus, product, or process disclosed herein, or represents that its use by such third party would not infringe privately owned rights. The views expressed herein are not necessarily those of the U.S. Nuclear Regulatory Commission.

INL/EXT-19-54609

Enhanced Component Performance Study: Emergency Diesel Generators

1998–2018

Zhegang Ma

Update Completed September 2019

Idaho National Laboratory Risk Assessment and Management Services Department

Idaho Falls, Idaho 83415

http://www.inl.gov

Prepared for the Division of Risk Assessment

Office of Nuclear Regulatory Research U.S. Nuclear Regulatory Commission

NRC Agreement Number NRC-HQ-60-14-D-0018

Enhanced Component Performance Study iii 2018 Update

Emergency Diesel Generators September 2019

ABSTRACT

This report presents an enhanced performance evaluation of emergency

diesel generators (EDGs) at U.S. commercial nuclear power plants. This report

evaluates component performance over time using (1) Institute of Nuclear Power

Operations (INPO) Consolidated Events Database (ICES) data from 1998

through 2018 and (2) maintenance unavailability (UA) performance data from

Mitigating Systems Performance Index (MSPI) Basis Document data from 2002

through 2018. The objective is to show estimates of current failure probabilities

and rates related to EDGs, trend these data on an annual basis, determine if the

current data are consistent with the probability distributions currently

recommended for use in NRC probabilistic risk assessments, show how the

reliability data differ for different EDG manufacturers and for EDGs with

different ratings; and summarize the subcomponents, causes, detection methods,

and recovery associated with each EDG failure mode. Engineering analyses

were performed with respect to time period and failure mode without regard to

the actual number of EDGs at each plant. The factors analyzed are: sub-

component, failure cause, detection method, recovery, manufacturer, and EDG

rating.

A statistically significant increasing trend was identified in the frequency of

FTLR demands for emergency power system (EPS) and high pressure core spray

(HPCS) EDGs and a statistically significant decreasing trend was identified in

the frequency of failure to run > 1H hours for EPS and HPCS EDGs.

Enhanced Component Performance Study iv 2018 Update


Enhanced Component Performance Study v 2018 Update


CONTENTS

ABSTRACT ................................................................................................................................................. iii

ACRONYMS ............................................................................................................................................... ix

1. INTRODUCTION ................................................................................................................................ 1

2. SUMMARY OF FINDINGS ................................................................................................................ 3 2.1 Increasing Trends ...................................................................................................................... 3

2.1.1 Extremely Statistically Significant ..................................................................................... 3 2.1.2 Highly Statistically Significant ........................................................................................... 3 2.1.3 Statistically Significant ....................................................................................................... 3

2.2 Decreasing Trends ..................................................................................................................... 3 2.2.1 Extremely Statistically Significant ..................................................................................... 3 2.2.2 Highly Statistically Significant ........................................................................................... 3 2.2.3 Statistically Significant ....................................................................................................... 3

2.3 Consistency Check Results ....................................................................................................... 3

3. FAILURE PROBABILITIES AND FAILURE RATES ...................................................................... 5 3.1 Overview ................................................................................................................................... 5 3.2 EDG Failure Probability and Failure Rate Trends .................................................................... 5

4. EDG UNAVAILABILITY AND REPAIR TIMES ............................................................................ 11 4.1 EDG Unavailability Overview ................................................................................................ 11 4.2 EDG Unavailability Trends ..................................................................................................... 11 4.3 Emergency Diesel Generator Repair Times ............................................................................ 13

5. EDG UNRELIABILITY TRENDS .................................................................................................... 15

6. ENGINEERING ANALYSIS ............................................................................................................. 17 6.1 Engineering Trends ................................................................................................................. 17 6.2 Comparison of ICES EPS EDG Unplanned Demand Results with Industry Results .............. 21 6.3 EPS EDG Performance by Manufacturer ................................................................................ 22 6.4 EPS EDG Performance by Rating ........................................................................................... 23 6.5 EPS EDG Engineering Analysis by Failure Modes ................................................................ 23

7. EPS EDG ASSEMBLY DESCRIPTION ........................................................................................... 32

8. DATA TABLES ................................................................................................................................. 33

9. REFERENCES .................................................................................................................................... 50

Enhanced Component Performance Study vi 2018 Update


FIGURES

Figure 1. Failure probability estimate trend for EPS EDG FTS. ................................................................. 8

Figure 2. Failure probability estimate trend for EPS EDG FTLR. .............................................................. 8

Figure 3. Failure rate estimate trend for EPS EDG FTR>1H. ..................................................................... 9

Figure 4. Failure probability estimate trend for HPCS EDG FTS. .............................................................. 9

Figure 5. Failure probability estimate trend for HPCS EDG FTLR. ......................................................... 10

Figure 6. Failure rate estimate trend for HPCS EDG FTR>1H. ................................................................ 10

Figure 7. EPS EDG UA trend. ................................................................................................................... 12

Figure 8. HPCS EDG UA trend. ................................................................................................................ 12

Figure 9. EPS EDG unreliability trend (8-hour mission). .......................................................................... 15

Figure 10. HPCS EDG unreliability trend (8-hour mission). ..................................................................... 16

Figure 11. Frequency of start demands (demands per reactor year) for EPS and HPCS EDGs. ............... 18

Figure 12. Frequency of FTLR demands (demands per reactor year) trend for EPS and HPCS

EDGs. ............................................................................................................................................... 18

Figure 13. Frequency of run > 1H hours (hours per reactor year) trend for EPS and HPCS EDGs. ......... 19

Figure 14. Frequency of FTS events (events per reactor year) trend for EPS and HPCS EDGs. .............. 19

Figure 15. Frequency of FTLR events (events per reactor year) trend for EPS and HPCS EDGs. ........... 20

Figure 16. Frequency of FTR>1H events (events per reactor year) trend for EPS and HPCS

EDGs. ............................................................................................................................................... 20

Figure 17. EPS EDG failure breakdown by sub component and failure mode .......................................... 26

Figure 18. EPS EDG breakdown by cause group and failure mode .......................................................... 27

Figure 19. EPS EDG component failure distribution failure mode and method of detection .................... 28

Figure 20. EPS EDG component failure distribution by failure mode and recovery determination ......... 29

Figure 21. EPS EDG failure distribution by manufacturer ........................................................................ 30

Figure 22. EPS EDG component failure modes by EPS EDG rating ........................................................ 31

Enhanced Component Performance Study vii 2018 Update


TABLES

Table 1. EDG systems. ................................................................................................................................. 5

Table 2. Industry-wide distributions of p (failure probability) and λ (hourly rate) for EPS EDGs,

from the 2015 Update. ........................................................................................................................ 5

Table 3. Industry-wide distributions of p (failure probability) and λ (hourly rate) for HPCS

EDGs. ................................................................................................................................................. 5

Table 4. Industry-average unavailability estimates for EPS EDGs. .......................................................... 11

Table 5. Weibull curve fit parameters. ....................................................................................................... 13

Table 6. Probability of exceeding selected EDG repair times (2009 – 2018). ........................................... 14

Table 7. Summary of EDG failure counts for the FTS failure mode over time by system. ....................... 21

Table 8. Summary of EDG failure counts for the FTLR failure mode over time by system. .................... 21

Table 9. Summary of EDG failure counts for the FTR > 1H failure mode over time by system. ............. 21

Table 10. EPS EDG unplanned demand performance comparison with industry-average

performance from ICES data. ........................................................................................................... 22

Table 11. EPS EDG manufacturer performance compared with industry-average performance—

FTS, FTLR, and FTR>1H combined................................................................................................ 22

Table 12. EPS EDG rating performance compared with industry-average performance—FTS,

FTLR, and FTR>1H combined. ....................................................................................................... 23

Table 13. Component failure cause groups. ............................................................................................... 24

Table 14. EPS EDG manufacturer population and total failure count. ...................................................... 30

Table 15. EPS EDG population by rating. ................................................................................................. 31

Table 16. Plot data for Figure 1, failure probability estimate trend for EPS EDG FTS ............................ 34

Table 17. Plot data for Figure 2, failure probability estimate trend for EPS EDG FTLR.......................... 35

Table 18. Plot data for Figure 3, failure rate estimate trend for EPS EDG FTR>1H ................................ 36

Table 19. Plot data for Figure 4, failure probability estimate trend for HPCS EDG FTS ......................... 37

Table 20. Plot data for Figure 5, failure probability estimate trend for HPCS EDG FTLR ...................... 38

Table 21. Plot data for Figure 6, failure rate estimate trend for HPCS EDG FTR>1H ............................. 39

Table 22. Plot data for Figure 7, EPS EDG UA trend ............................................................................... 40

Table 23. Plot data for Figure 8, HPCS EDG UA trend ............................................................................ 41

Enhanced Component Performance Study viii 2018 Update


Table 24. Plot data for Figure 9, EPS EDG unreliability trend (8-hour mission) ...................................... 42

Table 25. Plot data for Figure 10, HPCS EDG unreliability trend (8-hour mission) ................................. 43

Table 26. Plot data for Figure 11, frequency of start demands (demands per reactor year) trend

for EPS and HPCS EDGs ................................................................................................................. 44

Table 27. Plot data for Figure 12, frequency of FTLR demands (demands per reactor year) trend


Table 28. Plot data for Figure 13, frequency of run > 1H hours (hours per reactor year) trend for

EPS and HPCS EDG ........................................................................................................................ 46

Table 29. Plot data for Figure 14, frequency of FTS events (events per reactor year) trend for

EPS and HPCS EDGs ....................................................................................................................... 47

Table 30. Plot data for Figure 15, frequency of FTLR events (events per reactor year) trend for

EPS and HPCS EDGs ....................................................................................................................... 48

Table 31. Plot data for Figure 16, frequency of FTR>1H events (events per reactor year) trend


Enhanced Component Performance Study ix 2018 Update


ACRONYMS

AOV air-operated valve

CNID constrained non-informative prior distribution

CY calendar year

EDG emergency diesel generator

EPIX Equipment Performance and Information Exchange

EPS emergency power supply

ESF engineered safety feature

FTLR failure to load and run

FTR>1H failure to run > 1 hour

FTS failure to start

FY fiscal year

HPCS high-pressure core spray

ICES INPO Consolidated Events Database

INL Idaho National laboratory

INPO Institute of Nuclear Power Operations

IRIS Industry Reporting and Information System

MDP motor-driven pump

MOV motor-operated valve

MSPI Mitigating Systems Performance Index

NPRDS Nuclear Plant Reliability Data System

NRC Nuclear Regulatory Commission

OLS ordinary least squares

PRA probabilistic risk assessment

TDP turbine-driven pump

UA unavailability

Enhanced Component Performance Study x 2018 Update


Enhanced Component Performance Study 1 2018 Update


Enhanced Component Performance Study: Emergency Diesel Generators

1998–2018

1. INTRODUCTION

This report presents a performance evaluation of emergency diesel generators (EDGs) at U.S.

commercial nuclear power plants from 1998 through 2018. The objective is to show estimates of current

failure probabilities and rates related to EDGs, trend these data on an annual basis, determine if the

current data are consistent with the probability distributions currently recommended for use in NRC

probabilistic risk assessments, show how the reliability data differ for different EDG manufacturers and

for EDGs with different ratings; and summarize the subcomponents, causes, detection methods, and

recovery associated with each EDG failure mode. This year’s update continues with the three changes

implemented in the 2016 update that are different from prior updates: (1) the update results are based on

calendar year (CY) instead of the federal fiscal year (FY); (2) the failure events included in the update are

“hard” failures, i.e., the p-values indicating the likelihood the component would have failed during a 24-

hour mission are 1.0. Previous updates (2015 and before) include lesser p-values indicating a degraded

condition that probably would have caused failure during a 24-hour mission but were not quite hard

failures at their outset. (3) The discussion of EDG repair times, which was previously included in the

annual LOOP updates (see http://nrcoe.inl.gov/resultsdb/LOSP), is included in this report.

The enhanced component performance studies are conducted for the following component types: air-

operated valves (AOVs), EDGs, motor-driven pumps (MDPs), motor-operated valves (MOVs), and

turbine-driven pumps (TDPs). The EDG performance analysis was first conducted in 2007 with data from

1998 through 2006 [1] and then updated annually in a series of reports, with the last one being

documented in INL/LTD-17-44204, Enhanced Component Performance Study: Emergency Diesel

Generators 1998-2016 [2]. The Nuclear Regulatory Commission (NRC) Reactor Operational Experience

Results and Databases web page provides the links to the historical and current results of component

performance studies (http://nrcoe.inl.gov/resultsdb/CompPerf). An overview of the trending methods,

glossary of terms, and abbreviations is documented in the paper Overview and Reference [3] that can also

be found on that web page.

The data used in this study are based on the operating experience failure reports from Institute of

Nuclear Power Operations (INPO) Consolidated Events Database (ICES) [4], formerly the Equipment

Performance and Information Exchange Database (EPIX) and now upgraded again to IRIS, the Industry

Reporting and Information System. Maintenance unavailability (UA) performance data came from the

Reactor Oversight Process program’s Mitigating Systems Performance Index (MSPI) program [5] and

ICES. Previously, the study relied on operating experience obtained from licensee event reports, Nuclear

Plant Reliability Data System (NPRDS), and ICES. The ICES database, now IRIS, (which includes the

MSPI designated devices as a subset) has matured to the point where both component availability and

reliability can be estimated with a high degree of accuracy. In addition, the population of data in current

ICES database is much larger than the population available in the previous study.

The EDG failure modes considered are failure to start (FTS), failure to load and run (FTLR), and

failure to run greater than 1-hour (FTR>1H). Annual failure probabilities (failures per demand) are

provided for FTS and FTLR events and annual failure rates (failures per run hour) are provided for FTR >

1H events. EDG train maintenance unavailability probabilities are also considered. In addition to the

presentation of the component failure mode data and the UA data, an 8-hour component total unreliability

is calculated and trended. Each of the estimates is trended for the most recent 10-year period. Yearly

http://nrcoe.inl.gov/resultsdb

http://nrcoe.inl.gov/resultsdb/CompPerf

http://nrcoe.inl.gov/resultsdb/publicdocs/Overview-and-Reference.pdf



estimates have been provided for the entire study period. The results are reported separately for

emergency power system (EPS) and high pressure core spray (HPCS) EDGs.

While this report provides an overview of operational data and evaluate component performance over

time, it makes no attempt to estimate values for use in probabilistic risk assessments (PRAs). The 2015

Component Reliability Update [6], is an update to NUREG/CR-6928, Industry-Average Performance for

Components and Initiating Events at U.S Commercial Nuclear Power Plants [7], and reports the TDP

unreliability estimates for probabilistic risk assessments. Estimates from that report are included herein,

for comparison. Those estimates are labelled “2015 Update” (or “Update 2015”) in the associated tables

and figures.

Section 2 of this report presents the summary of findings from the study, with particular interest in the

existence of any statistically significant increasing or decreasing trends in component performances.

Section 3 provides annual estimates of failure probabilities and rates related to EDGs as well as the

trending of the estimates. Section 4 provides EDG train UA estimates and their trends (Sections 4.1 and

4.2), as well as the discussion of EDG repair times (Section 4.3). Section 5 estimates the annual total

unreliability and the trends for EDG.

Section 6 presents various engineering analyses performed for EDG with respect to time period and

failure modes. In Section 6.1, the same failures used in Section 3 are used to compute estimates of overall

failure frequencies per plant reactor year (with EPS and HPCS EDG failures combined). Frequencies of

demands per plant reactor year are also provided for each year, for each of the possible failure modes. As

in Section 3, each of the estimates is trended for the most recent 10-year period. The frequencies show

general industry performances and are not based on the number of EDGs at each plant.

In Sections 6.2 through 6.4, various subsets of the EDG data are compared with the distributions

currently recommended for PRA use in the “2015 Update.” First, the subset of failure events and

demands from this report that occurred on unplanned demands (engineered safety feature actuations) is

compared for consistency with the 2015 Update data. This evaluation provided a check on the ongoing

use of the 2015 Update EDG data (which includes failures from possibly incomplete testing demands). In

Section 6.3, data from each EDG manufacturer is compared. Finally, in Section 6.4, EDG failure

groupings based on EDG ratings are compared.

Section 6.5 provides breakdowns of the failures for each failure mode for the two plant systems with

EDGs. The analyses are based on the following factors: sub-component, failure cause, detection method,

and recovery.

Section 7 provides additional information on the EDG assembly component boundaries and failure

modes. Section 8 presents the plot data for various figures in previous sections.

http://nrcoe.inl.gov/resultsdb/publicdocs/AvgPerf/ComponentReliabilityDataSheets2010.pdf




2. SUMMARY OF FINDINGS

The results of this study are summarized in this section. Of particular interest is the existence of any

statistically significanta increasing trends.

2.1 Increasing Trends

2.1.1 Extremely Statistically Significant

None.

2.1.2 Highly Statistically Significant

None.

2.1.3 Statistically Significant

A statistically significant increasing trend in the frequency of FTLR demands for EPS

and HPCS EDGs was identified with a p-value of 0.0165 (see Figure 12). This is a new

trend that was not observed in the 2016 EDG update study [2].

2.2 Decreasing Trends

2.2.1 Extremely Statistically Significant

None.

2.2.2 Highly Statistically Significant

None.

2.2.3 Statistically Significant

A statistically significant decreasing trend in the frequency of run > 1H hours for EPS

and HPCS EDGs was identified with a p-value of 0.0198 (see Figure 13). This trend had

also been observed in the 2016 EDG update study as being highly statistically significant.

2.3 Consistency Check Results

An ongoing concern in the nuclear risk assessment field is whether industry failure rate estimates that

are largely derived from test data adequately predict component performance during unplanned (ESF)

demands. Section 6.2 provides results of a consistency check that compare industry failure estimates

obtained via simulation test on parameters from the 2015 Update against operational failure counts

obtained from actual EDG performance with ESF demands. These consistency checks show that the FTS,

FTLR, and FTR failure observations in the non-test, operational ESF demand data lie within their

corresponding industry-average failure estimate distributions, provided in the 2015 Update (Table 2), that

were based on both test and non-test operational ESF demands.

a. Statistically significant is defined in terms of the ‘p-value.’ A p-value is a probability indicating whether to accept

or reject the null hypothesis that there is no trend in the data. P-values of less than or equal to 0.05 indicate that we

are 95% confident that there is a trend in the data (reject the null hypothesis of no trend.) By convention, we use the

"Michelin Guide" scale: p-value < 0.05 (statistically significant), p-value < 0.01 (highly statistically significant); p-

value < 0.001 (extremely statistically significant).



Section 6.3 provides the results of consistency checks by EDG manufacturer. Two manufacturer’s

ESF EPS EDG failure counts lie in the upper 95% of the uncertainty range of the industry-average

estimate. However, these manufacturers have very few EPS EDGs, and so the data are limited. The

remaining manufacturers’ failure counts lie within the 5% to 95% interval of the industry-average

estimate uncertainty band.

Section 6.4 shows the results of the consistency check by EDG load rating. The failure counts by

rating all lie within the 5% to 95% interval of the industry-average estimate uncertainty band.



3. FAILURE PROBABILITIES AND FAILURE RATES

3.1 Overview

The failure probabilities and failure rates of EDGs have been calculated from the operating

experience for FTS, FTLR, and FTR>1H. The EDG data set obtained from ICES includes EDGs in the

systems listed in Table 1. Table 2 shows failure probability and failure rate estimates for the EPS EDG

from Reference [6], or the 2015 Update. Table 3 shows the failure probability and failure rate estimates

for the HPCS EDG. The HPCS EDG failure probability was not fully analyzed in [6] and is presented

here based on the current ICES data that has been reviewed at Idaho National Laboratory (INL).

Table 1. EDG systems.

System Description EDG Count

EPS Emergency power supply 232 HPCS High pressure core spray 8 Total 240

The EDGs do not operate all the time. They are standby-components required to operate when called

upon, both when the reactor is critical and during shutdown periods. The demands and run hours are

reported on a quarterly or semi-annual basis through the MSPI program. All demand types are

considered—testing, non-testing, and those ESF demands that require the EDG to mitigate a bus under-

voltage condition.

Table 2. 2015 Update industry-wide distributions of p (failure probability) and λ (hourly rate) for EPS

EDG.

Failure Mode 5% Median Mean 95%

Distribution

Type

FTS 1.45E-3 2.73E-3 2.83E-3 4.59E-3 Beta 8.59 3.02E+03 FTLR 1.18E-3 3.38E-3 3.73E-3 7.42E-3 Gamma 3.61 9.70E+02 FTR>1H 3.78E-4 1.35E-3 1.54E-3 3.34E-3 Gamma 2.68 1.74E+03

Table 3. Industry-wide distributions of p (failure probability) and λ (hourly rate) for HPCS EDGs.

Failure Mode 5% Median Mean 95%

Distribution

Type

FTS 2.19E-4 8.33E-4 9.56E-4 2.12E-3 Beta 2.50 2.61E+03 FTR 4.25E-4 1.07E-3 1.15E-3 2.16E-3 Gamma 4.50 3.91E+03

3.2 EDG Failure Probability and Failure Rate Trends

This section estimates industry-wide annual failure probabilities and failure rates for EDGs in the

entire study period which covers 1998 through 2018. The estimates are trended for the most recent 10-

year period.

The failure probability and failure rate estimates in this section were obtained from a Bayesian update

process. The means from the posterior distributions were plotted for each year. The 5th and 95th

percentiles from the posterior distributions are also provided and give an indication of the relative

uncertainty in the estimated parameters from year to year. When there are no failures, the interval is

larger than the interval for years when there are one or more failures because of the form of the posterior

variance. Each update utilizes a relatively “flat” constrained non-informative prior distribution (CNID),

which has wide bounds, see [3] and NUREG/CR-6823 [8]. CNID is a compromise between an



informative prior and the Jeffreys noninformative prior. The mean of the CNID uses prior belief and is

based on a pooling of the component or event type data for the years going into the plot (i.e., the most

recent 10-year period), but the dispersion is defined to correspond to little information (i.e., relatively flat

by set) so that the prior distributions did not create large changes in the data.

For failure rates or Poisson data, the CNID is a gamma distribution, with the mean (𝜇) given by

prior belief and calculated as:

𝜇 =∑ 𝑓𝑖 + 0.5

∑ 𝑇𝑖

(1)

where 𝑓𝑖 and 𝑇𝑖 are the failures and operating/standby time for the ith year, respectively. The CNID shape

parameter = 0.5. The posterior distribution mean for the ith year (𝜇𝑖) can be calculated as:

𝜇𝑖 =𝑓𝑖 + 0.5

0.5𝜇

+ 𝑇𝑖

(2)

For failure probabilities or binomial data, the CNID is a beta approximation, with the mean given by

prior belief and calculated as:

𝜇 =∑ 𝑓𝑖 + 0.5

∑ 𝐷𝑖 + 1

(3)

where 𝑓𝑖 and 𝐷𝑖 are the failures and demands for the ith year, respectively. The CNID shape parameter (𝛼)

is a number between 0.3 and 0.5 based on the mean μ (see Table C.8 of [8]). The posterior distribution

mean for the ith year (𝜇𝑖) can be calculated as:

𝜇𝑖 =𝑓𝑖 + 𝛼𝛼𝜇 + 𝐷𝑖

(4)

The horizontal curves plotted around the regression lines in the graphs form 90% simultaneous

confidence bands for the fitted lines. The bounds are larger than ordinary confidence bands for the

individual coefficients because they form a confidence band for the entire line. In the lower left hand

corner of the trend figures, the regression p-values are reported. They come from a statistical test to

assess evidence against the slope of the regression line being zero. Low p-values indicate strong evidence

that the slopes are not zero, and suggest a trend does exist. P-values of less than or equal to 0.05 indicate

that we are 95% confident that there is a trend in the data (reject the null hypothesis of no trend.) By

convention, this study uses the "Michelin Guide" scale: p-value < 0.05 (statistically significant), p-value <

0.01 (highly statistically significant); p-value < 0.001 (extremely statistically significant).

The regression methods are all based on “ordinary least squares” (OLS), which minimizes the

residuals, or the square of the vertical distance between the annual data points and the fitted regression

line. The p-values assume normal distributions for the residuals, with the same variability in the residuals

across the years. In the case where the data involve failure counts, the iterative reweighted least squares

is used to account for the fact that count data are not expected to have a constant variance (for example,

the variance for Poisson-distributed counts is equal to the expected number of counts, which is expected

to vary proportionally to the expected number of counts). Further information on the trending methods is

provided in Section 2 of the Overview and Reference document [3].

A final feature of the trend graphs is the 2015 Update baseline industry values from Table 2 which are

shown for comparison.

Figure 1 to Figure 6 provide the plots for industry-wide failure probabilities/rates of EPS and HPCS

EDGs. The data for these plots are provided in Section 8.




o Figure 1 shows the failure probability estimate trends for EPS EDG FTS.

o Figure 2 shows the failure probability estimate trends for EPS EDG FTLR.

o Figure 3 shows the failure rate estimate trends for EPS EDG FTR>1H.

o Figure 4 shows the failure probability estimate trends for HPCS EDG FTS.

o Figure 5 shows the failure probability estimate trends for HPCS EDG FTLR.

o Figure 6 shows the failure rate estimate trends for HPCS EDG FTR>1H.

No trends were identified for EPS or HPCS EDG failure probabilities/rates for FTS, FTLR, and

FTR>1H events in the most recent 10-year period.



Figure 1. Failure probability estimate trend for EPS EDG FTS.

Figure 2. Failure probability estimate trend for EPS EDG FTLR.



Figure 3. Failure rate estimate trend for EPS EDG FTR>1H.

Figure 4. Failure probability estimate trend for HPCS EDG FTS.



Figure 5. Failure probability estimate trend for HPCS EDG FTLR.

Figure 6. Failure rate estimate trend for HPCS EDG FTR>1H.



4. EDG UNAVAILABILITY AND REPAIR TIMES

4.1 EDG Unavailability Overview

The industry-wide test or maintenance UA of EDG trains has been calculated from operating

experience. UA data are for EDG trains, which can include more than just the EDG. However, in most

cases the EDG contributes the majority of the UA reported. Table 4 shows overall results for the EDG

from the 2015 Update [6] which based on UA data from the MSPI program and ICES. In the

calculations, planned and unplanned unavailable hours for a train are combined.

Table 4. 2015 Update industry-average unavailability estimates for EPS EDGs.

Description Distribution Mean

EDG Test or Maintenance (EPS) Normal 1.48E-2 0.0148 0.0063

EDG Test or Maintenance (HPCS) Normal 1.17E-2 0.0117 0.0025

4.2 EDG Unavailability Trends

The graphs that follow provide overall maintenance unavailability, planned (such as test) and

unplanned, data for the 1998–2018 period. Note that these data do not supersede the data in Table 4 for

use in risk assessments.

Trends in EDG train unavailability are shown in Figure 7 and Figure 8. Data tables for these figures

are presented in Section 8. The EDGs in systems EPS and HPCS are trended. The yearly unavailability

and reactor critical hour data were obtained from the Reactor Oversight Process program (1998 to 2001)

and MSPI EPS indicator (2002 to 2018). The total EDG downtimes during operation for each plant and

year were summed and divided by the corresponding number of EDG-reactor critical hours.

Unavailability data for plant shutdown periods are not reported.

A change in reporting requirements for UA occurred in 2002. The Reactor Oversight Process

program data (1998–2001) did not include EDG overhaul outages while plants were in critical operation,

while the MSPI (2002–2018) requires plants to report such outages. The difference in the annual means

of these two groups is not statistically significant.

The mean and variance for each year is the sample mean and variance calculated from the plant-level

unavailability’s for that year. The vertical bar spans the calculated 5th to 95th percentiles of the beta

distribution with matching means.

For the trend graphs, a least squares fit is sought for the linear or logit model. Section 3 in the

Overview and Reference document provides further information [3]. In the lower left hand corner of the

trend figures, the p-value is reported. A review of the p-values identified no statistically significant trends

in EDG unavailability for the most recent 10-year period.



Figure 7. EPS EDG UA trend.

Figure 8. HPCS EDG UA trend.



4.3 Emergency Diesel Generator Repair Times

The data for repair times performed under actual emergency conditions are not available so repair

durations were based on the number of hours of unplanned unavailability that have been reported for each

EDG from 2009 to 2018. The hourly unplanned unavailability is reported to the NRC in the MSPI data.

This MSPI data were not reported prior to 2006.

A Weibull distribution was fitted to the unplanned unavailability durations. The Weibull fit

parameters are provided in Table 5. The probability an EDG unplanned outage duration exceeds a given

time (t) is listed in Table 6. The correspondence between fitted and observed distributions is very good at

short to moderate times but not as good at very long repair times (well beyond typical PRA missions),

such as when the outage spans hundreds of hours. In that situation, the long right tail of the repair time

distribution is fit better by a lognormal distribution than a Weibull.

Table 5. Weibull curve fit parameters.

Parameter Value

Mean 34.2

Median 15.4

Weibull(α) 0.681

Weibull(β) 26.3



Table 6. Probability of exceeding selected EDG repair times (2009 – 2018).

Recovery Time (hr)

Weibull Model

Probability

0.5 0.935

1 0.898

1.5 0.868

2 0.841

3 0.796

4 0.758

5 0.724

6 0.694

7 0.667

8 0.641

9 0.618

10 0.596

11 0.576

12 0.557

13 0.539

14 0.522

15 0.506

16 0.491

17 0.476

18 0.462

19 0.449

20 0.436

21 0.424

22 0.413

23 0.402

24 0.391



5. EDG UNRELIABILITY TRENDS

Trends in total component unreliability are shown in Figure 9 and Figure 10. Plot data for these

figures are provided in Section 8. Total unreliability is defined as the union of UA, FTS, FTLR,

FTR>1H. The probability of FTR>1H is calculated for 7 hours to provide the results for an 8-hour

mission. The trends are shown at the system-specific level across the industry. The trending method is

described in more detail in Section 4 of the Overview and Reference document [3]. In the lower left hand

corner of the trend figures, the regression method and the p-value are reported. A review of the p-values

identifies no statistically significant trends in the EDG total unreliability estimates for the most recent 10-

year period.

No “2015 Update” data for use in risk assessments are cited for EDG unreliability because these data

are generated a-priori. The risk assessment models compute unreliability as an output rather than an

input.

Figure 9. EPS EDG unreliability trend (8-hour mission).



Figure 10. HPCS EDG unreliability trend (8-hour mission).



6. ENGINEERING ANALYSIS

This section present the engineering analysis for factors that could influence the system and

component trends. Engineering trends of component failures and demands are presented in Section 6.1.

Differences between testing and actual unplanned demands are considered in Section 6.2. Differences

among manufacturers are presented in Section 6.3, and differences among EDG ratings are presented in

Section 6.4. Finally, engineering analyses performed with respect to failure mode are presented in

Section 6.5. The failure mode factors analyzed were: sub-component, failure cause, detection method,

manufacturer, and EDG rating.

6.1 Engineering Trends

This section presents frequency trends for EPS and HPCS EDG failures and demands. The data are

normalized by reactor year for plants that report data for these EDGs, with no consideration for plant

system (EPS versus HPCS) or for the number of EDGs at a plant. The trends provide an overview of the

demand counts and failure counts associated with each failure mode across the years.

o Figure 11 shows the trend for EPS and HPCS EDG frequency of start demands.

o Figure 12 shows the trend for EPS and HPCS EDG frequency of load and run demands.

o Figure 13 shows the trend for the EPS and HPCS EDG frequency of run hours.

o Figure 14 shows the trend for EPS and HPCS EDG frequency of FTS events.

o Figure 15 shows the trend EPS and HPCS EDG frequency of FTLR events.

o Figure 16 shows the trend for the EPS and HPCS EDG frequency of FTR>1H events.

The data for the above figures are provided in Section 8. The standby systems from Table 1 are

trended together for each figure. The rate methods described in Section 2 of the Overview and Reference

document are used [6].

In the lower left hand corner of the above trend figure, the regression p-values are reported. A review

of these p-values identified the following trend for the most recent 10-year period:

o Statistically significant increasing trend in the frequency of FTLR demands for EPS and

HPCS EDGs, with a p-value of 0.0165 (see Figure 12). This is a new trend that was not

observed in the 2016 EDG update study [2].

o Statistically significant decreasing trend in the frequency of run > 1H hours for EPS and

HPCS EDGs, with a p-value of 0.0198 (see Figure 13). This trend has been observed in the

2016 EDG update study as highly statistically significant.

Table 7–Table 9 provide a summary of the total failure event count for each of the years for which a

trend line is plotted. Table 7 summarizes the failures by system and year for the FTS failure mode.

Table 8 summarizes the failures by system and year for the FTLR failure mode. Table 9 summarizes the

failures by system and year for the FTR>1H failure mode. The data in Table 7–Table 9 show failure

events resulting from FTLR and FTR>1H occur in roughly equal numbers, while FTS failures occur

somewhat less frequently than FTLR and FTR>1H failures. Furthermore, HPCS EDGs are about

3percent of the EDG population, but account for only 1 to 2 percent of the failure counts throughout the

period being trended.




Figure 11. Frequency of start demands (demands per reactor year) for EPS and HPCS EDGs.

Figure 12. Frequency of FTLR demands (demands per reactor year) trend for EPS and HPCS EDGs.



Figure 13. Frequency of run > 1H hours (hours per reactor year) trend for EPS and HPCS EDGs.

Figure 14. Frequency of FTS events (events per reactor year) trend for EPS and HPCS EDGs.



Figure 15. Frequency of FTLR events (events per reactor year) trend for EPS and HPCS EDGs.

Figure 16. Frequency of FTR>1H events (events per reactor year) trend for EPS and HPCS EDGs.



Table 7. Summary of EDG failure counts for the FTS failure mode over time by system.

System EDG

Count EDG

Percent

Year

Total

Percent of

Failures 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018

EPS 232 96.7 % 15 15 19 14 6 12 13 13 10 7 124 98.4 %

HPCS 8 3.3 % 1 1 2 1.6 %

Total 240 100% 15 15 19 14 6 13 13 13 10 8 126 100%

Table 8. Summary of EDG failure counts for the FTLR failure mode over time by system.

System EDG

Count EDG

Percent

Year

Total

Percent of

Failures 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018

EPS 232 96.7 % 18 11 16 17 12 12 13 11 9 13 132 96.4 %

HPCS 8 3.3 % 1 1 1 1 1 5 3.6 %

Total 240 100% 18 12 16 18 12 13 14 11 9 14 137 100%

Table 9. Summary of EDG failure counts for the FTR > 1H failure mode over time by system.

System EDG

Count EDG

Percent

Year

Total

Percent of

Failures 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018

EPS 232 96.7 % 8 13 21 11 17 17 12 10 22 10 141 96.6 %

HPCS 8 3.3 % 1 1 1 2 5 3.4 %

Total 240 100% 8 14 21 11 18 17 12 11 22 12 146 100%

6.2 Comparison of ICES EPS EDG Unplanned Demand Results with Industry Results

Because the ICES EPS EDG data are dominated by test demands (over 95% of the demands are

typically from tests), an ongoing concern is whether these mostly test data adequately predict EPS EDG

performance during unplanned demands. This comparison evaluates the same dataset for standby

components that is used for the overall trends shown in this document, but limits the failure data to those

that are discovered during an ESF demand that was reported in ICES. The data are further limited to

2003 to present since the ESF demand reporting in ICES is inconsistent prior to 2003.

To answer this question, ICES failure records were reviewed to identify actual unplanned EPS EDG

demands involving bus under voltage conditions. Such events require the associated EPS EDG to start,

load onto the bus and power the bus until normal power is recovered to the bus. There are additional EPS

EDG unplanned demands in which a bus under voltage condition did not exist. In those cases, the EPS

EDG did not have to load and power the bus. Such unplanned demands do not fully exercise the mission

of the EPS EDGs and therefore were not counted.

The EPS EDG unplanned demand data covering 2003 – 2018 are summarized in Table 10.

Consistency between the unplanned demand data and 2015 Update industry-average performance (from

Table 2) was evaluated using the predictive distribution approach outlined in the Handbook of Parameter

Estimation for Probabilistic Risk Assessment, NUREG/CR-6823, Sections 6.2.3.5 and 6.3.3.4 [8].

The unplanned demand data were aggregated at the industry level (failures and demands). The

industry-average failure mode distribution (from Table 2) was sampled and the predicted number of

events was evaluated using the binomial distribution with industry-average failure probability and

associated number of demands. This process was repeated 1000 times, each time obtaining the total

number of failures predicted by the industry average failure parameters. Then the actual number of

observed unplanned demand failures (listed in the “Observed Failures” column of Table 10) was



compared with this sample to determine the probability of observing this number of failures or greater. If

the probability was greater than 0.05 and less than 0.95, then the Table 2 industry-average distribution

obtained from the ICES data analysis is considered to be consistent with the observed unplanned demand

performance.

Table 10. EPS EDG unplanned demand performance comparison with industry-average performance

from ICES data.

Failure Modes Plants

Demands or Hours

Observed

Failures Expected Failures

Probability of

≥ Failures

Consistent with Industry-Average

Performance?

FTS 95 534 0 1.5 1.00 Yesa

FTLR 95 311 1 1.2 0.64 Yes

FTR 95 4637 4 7.1 0.74 Yes

a. In this case P(X=0) = 0.25 which is considered consistent with the industry average data.

These consistency checks show that the FTS, FTLR, and FTR failure observations in the non-test,

operational ESF demand data lie within their corresponding industry-average estimate distributions,

provided in the 2015 Update (Table 2), that were based on both test and non-test operational ESF

demands.

6.3 EPS EDG Performance by Manufacturer

Table 11 presents the results of summarizing EPS EDG performance by manufacturer. ICES contains

information on EPS EDG manufacturers, but over the years some manufacturers have changed names or

have been acquired by others. Therefore, in order to identify the original manufacturer, the ICES

information was supplemented by other EPS EDG reports. The results are a second consistency check

against the industry-average distributions in Table 2. The comparison was made for the combination of

all three failure modes.

Two manufacturer’s EPS EDG failure observations lie in the upper 95% of the uncertainty range of

the industry average distribution. However, these two manufacturers involve very few EPS EDGs. The

rest of the manufacturers’ failure observations lie within the 5% to 95% interval.

Table 11. EPS EDG manufacturer performance compared with industry-average performance— FTS,

FTLR, and FTR>1H combined.

Manufacturer Code EPS

EDGs Observed Failures

Expected Failures

Probability ≥ Observed

Failures


Performance?a

ALCO Power AP 24 69 88.4 0.74 Yes

Cooper Bessemer CB 37 75 135.5 0.93 Yes

Electro Motive/General Motors

EM/GM 69 203 248.2 0.70 Yes

Fairbanks Morse/Colt FM/C 67 250 250.2 0.46 Yes

Nordberg NB 8 37 36.5 0.44 Yes

SAC/Compair Luchard/ Jeumont Schndr

SC/JS 3 22 9.6 0.02 No

TransAmerica DeLaval TD 20 84 78.3 0.36 Yes

Worthington Corp WC 4 37 12.6 0.00 No



a. If the probability of observing the failures or greater is ≥ 0.05 and ≤ 0.95, then the industry-average estimate is considered consistent with the observed failure count.

6.4 EPS EDG Performance by Rating

Table 12 presents the results of the evaluation of EPS EDG performance by rating. The results are a

consistency check of the industry-average distributions in Table 2 against observed performance by EDG

rating. The comparison was made for the combination of all three failure modes. The failure

observations for ratings all lie within the 5% to 95% interval of the industry average distribution and are

therefore consistent with the industry-average failure rate estimates.

Table 12. EPS EDG rating performance compared with industry-average performance—FTS, FTLR, and

FTR>1H combined.

Rating EPS

EDGs Observed Failures

Expected Failures

Probability ≥

Observed Failures


Performance?a

450-249 KW 2 6 10.3 0.84 Yes

1,000-4,999 KW 170 576 631.3 0.57 Yes

5,000-99,999 KW 58 193 210.8 0.57 Yes

100,000-499,999 KW 2 2 4.8 0.92 Yes

a. If the probability of observing the actual failures or greater is ≥ 0.05 and ≤ 0.95, then the industry-average estimate is considered consistent with the observed failure count.

6.5 EPS EDG Engineering Analysis by Failure Modes

The engineering analysis of EPS EDG failure sub-components, causes, detection methods, and

recovery are presented in this section (There are too few HPCS EDGs to perform similar analyses on

them). The events are also categorized by the failure mode determined after ICES data review by the INL

staff. See Section 7 for more description of failure modes.

EPS EDG sub-component contributions to the three failure modes are presented in Figure 17. The

sub-component contributions are similar to those used in the CCF database. For FTS, the instrumentation

and control, generator, starting air, and engine piece parts are the top contributors to failures. For FTLR,

the instrumentation and control and breaker are dominating contributors to failures. For FTR, high

contributors include the cooling, engine, fuel oil, and instrumentation and control.

EPS EDG failure cause group contributions to the three failure modes are presented in Figure 18.

The cause groups have been re-arranged in this update study in order to align with those currently used in

the CCF database. Table 13 shows the breakdown of the cause groups with the specific causes that were

coded during the data collection. The most likely causes are component issues and human errors. The

Component cause group includes the causes that were related to something internal to the component or

an aging or worn out part, which were categorized as the Internal cause group in previous studies [2]. The

Human cause group is primarily influenced by maintenance and operating procedures and practices.

EPS EDG failure detection methods for the three failure modes are presented in Figure 19. The

most likely detection method is testing, which is the prevalent detection method for most standby

components. The inspection failure detection method is also important in the FTS failure mode.



EPS EDG recovery results for the three failure modes are presented in Figure 20. Most EPS EDG

failures were judged to not be recoverable. The overall non-recovery to recovery ratio is approximately

14:1, meaning that 14 of every 15 failures were not recovered.

Table 13. Component failure cause groups. a

Group Specific Cause Description

Component Internal to component, piece-part

Used when the cause of a failure is a non-specific result of a failure internal to the component that failed other than aging or wear.

Set point drift Used when the cause of a failure is the result of set point drift or adjustment.

Age/Wear Used when the cause of the failure is a non-specific aging or wear issue.

Design Construction/installation error or inadequacy

Used when a construction or installation error is made during the original or modification installation. This includes specification of incorrect component or material.

Design error or inadequacy

Used when a design error is made.

Manufacturing error or inadequacy

Used when a manufacturing error is made during component manufacture.

Environment Ambient environmental stress

Used when the cause of a failure is the result of an environmental condition from the location of the component.

Internal environment The internal environment led to the failure. Debris/Foreign material as well as an operating medium chemistry issue.

Extreme environmental stress

Used when the cause of a failure is the result of an environmental condition that places a higher than expected load on the equipment and is transitory in nature.

Human Accidental action (unintentional or undesired human errors)

Used when a human error (during the performance of an activity) results in an unintentional or undesired action.

Human action procedure Used when the correct procedure is not followed or the wrong procedure is followed. For example: when a missed step or incorrect step in a surveillance procedure results in a component failure.

Inadequate maintenance Used when a human error (during the performance of maintenance) results in an unintentional or undesired action.

Inadequate procedure Used when the cause of a failure is the result of an inadequate procedure operating or maintenance.

Other State of other component Used when the cause of a failure is the result of a component state that is not associated with the component that failed. An example would be the diesel failed due to empty fuel storage tanks.

Other (stated cause does not fit other categories)

Used when the cause of a failure is provided but it does not meet any one of the descriptions.

a . The cause groups have been re-arranged to align with those currently used in the CCF database.



Unknown Used when the cause of the failure is not known.



Figure 17. EPS EDG failure breakdown by sub component and failure mode



Figure 18. EPS EDG breakdown by cause group and failure mode



Figure 19. EPS EDG component failure distribution failure mode and method of detection



Figure 20. EPS EDG component failure distribution by failure mode and recovery determination



Figure 21 shows the percentage of failure events for the three failure modes segregated by EPS EDG

manufacturer as indicated in the ICES database. Table 14 shows the distribution of the various

manufacturers of EPS EDGs in the ICES database and the total failure count associated with each. Based

on the information given in Figure 21, EPS EDG manufacturers do not appear to be correlated to any

particular failure mode pattern.

Figure 21. EPS EDG failure distribution by manufacturer

Table 14. EPS EDG manufacturer population and total failure count.

Manufacturer Code EPS

EDGs

% of EDGs

Total Failure Count

% of Failures

ALCO Power AP 24 10.3% 69 8.9%

Cooper Bessemer CB 37 15.9% 75 9.7%

Electro Motive EM/GM 69 29.7% 203 26.1%

Fairbanks Morse/Colt FM/C 67 28.9% 250 32.2%

Nordberg NB 8 3.4% 37 4.8%

SAC/Compair Luchard/Jeumont Schndr SC/JS 3 1.3% 22 2.8%

Transamerica Delaval TD 20 8.6% 84 10.8%

Worthington Corp WC 4 1.7% 37 4.8%

Totals 232 100% 777 100%



Figure 22 shows the percentage of failure events for the three failure modes segregated by EPS EDG

rating as indicated in the ICES database. Table 15 shows the distribution of the various rated EPS EDGs

in the ICES database used in this study. The larger EDG differs from the others in not yet having any

FTS events, but the operational experience for this EDG is much shorter than for other EDGs.

Figure 22. EPS EDG component failure modes by EPS EDG rating

Table 15. EPS EDG population by rating.

EPS EDG Rating Device Count

% of Devices

Total Failure Count

% of Failures

50-249 KW 2 0.9% 6 0.8%

1,000-4,999 KW 170 73.3% 576 74.1%

5,000-99,999 KW 58 25.0% 193 24.8%

100,000-499,999 KW 2 0.9% 2 0.3%

Total 232 100% 777 100%



7. EPS EDG ASSEMBLY DESCRIPTION

The EDGs are those within the Class 1E ac electrical power system at U.S. commercial nuclear power

plants and those in the HPCS systems. Station blackout EDGs are not included.

The EDG includes the diesel engine with all components in the exhaust path, electrical generator,

generator exciter, output breaker, combustion air, lube oil systems, fuel oil system, and starting

compressed air system, and local instrumentation and control circuitry. The sequencer is excluded from

the EDG component. For the service water system providing cooling to the EDGs, only the devices

providing control of cooling flow to the EDG heat exchangers are included. Room heating and

ventilating is not included.

The EDG failure modes include FTS, FTLR, and FTR>1H. These failure modes were used in

NUREG/CR-6928 and are similar to those used in the MSPI Program. There is some uncertainty

concerning when the run hours should start to be counted; for example, should they start as soon as the

EDG starts or should they start only after the output circuit breaker has closed? For this study, the run

hours start as soon as the EDG is started, which is the way data have been reported in ICES. The total run

hours are partitioned by failure mode, with the first hour being used for FTLR, and the remaining hours

assigned to FTR>1H.

Guidelines for determining whether a component failure event reported in ICES is to be included in

FTS, FTLR, or FTR>1H are similar to those used in the MSPI Program. In general, any circumstance in

which the component is not able to meet the performance requirements defined in the PRA is counted.

This includes conditions revealed through testing, operational demands, unplanned demands, or discovery

(see INPO 12-009 [9] for examples of operational demands, or operational non-test demands). Also, run

failures that occur beyond the typical 24-hour mission time in PRAs are included. However, certain

events are excluded: slow engine starting times that do not exceed the PRA success criteria, conditions

that are annunciated immediately in the control room without a demand, and run events representing

degraded conditions that are shown to not have caused an actual run failure within 24 hours. Also, events

occurring during maintenance or post-maintenance testing that are related to the actual maintenance

activities are excluded. Finally, in contrast to the MSPI Program, a general guideline on slow starting

times is to include only those slow starts requiring more than 20 seconds as FTS events, similar to what

was done for the CCF database and the EPS system study. (In the MSPI Program, most licensees chose

to use technical specification requirements for fast starts as their success criteria – typically less than 10

seconds to start.) All of the EDG events within ICES were reviewed to ensure that they were binned to

the correct failure mode—FTS, FTLR, FTR>1H, or no failure. However, even given detailed

descriptions of failure events, this binning still required some judgment and involves some uncertainty.

Guidelines for counting demands and run hours are similar to those in the MSPI Program. Start and

load/run demands include those resulting from tests, operational demands, and unplanned demands.

Demands during maintenance and post-maintenance testing are excluded. Similarly, run hours include

those from tests, operational demands, and unplanned demands. Note that the test demands and run hours

dominate the totals, compared with operational and unplanned demands and run hours.



8. DATA TABLES

In this section, the plot data for Figure 1 to Figure 16 in previous sections are provided in Table 16 to

Table 31, respectively.

Figure Table Analysis

Figure 1 Table 16 Failure probability estimate trend for EPS EDG FTS

Figure 2 Table 17 Failure probability estimate trend for EPS EDG FTLR

Figure 3 Table 18 Failure rate estimate trend for EPS EDG FTR>1H

Figure 4 Table 19 Failure probability estimate trend for HPCS EDG FTS

Figure 5 Table 20 Failure probability estimate trend for HPCS EDG FTLR

Figure 6 Table 21 Failure rate estimate trend for HPCS EDG FTR>1H

Figure 7 Table 22 EPS EDG UA trend

Figure 8 Table 23 HPCS EDG UA trend

Figure 9 Table 24 EPS EDG unreliability trend (8-hour mission)

Figure 10 Table 25 HPCS EDG unreliability trend (8-hour mission)

Figure 11 Table 26 Frequency of start demands (demands per reactor year) trendfor EPS and HPCS EDGs

Figure 12 Table 27 Frequency of FTLR demands (demands per reactor year) trend for EPS and HPCS EDGs

Figure 13 Table 28 Frequency of run > 1H hours (hours per reactor year) trend for EPS and HPCS EDG

Figure 14 Table 29 Frequency of FTS events (events per reactor year) trend for EPS and HPCS EDGs

Figure 15 Table 30 Frequency of FTLR events (events per reactor year) trend for EPS and HPCS EDGs

Figure 16 Table 31 Frequency of FTR>1H events (events per reactor year) trend for EPS and HPCS EDGs



Table 16. Plot data for Figure 1, failure probability estimate trend for EPS EDG FTS

Year Failures Demands

Regression Curve Data Points Plot Trend Error Bar Points

Mean Lower (5%)

Upper (95%)

Lower (5%)

Upper (95%) Mean

2015 Update 1.45E-03 4.59E-03 2.83E-03

1998 18 4,285 -- -- -- 2.71E-03 6.13E-03 4.16E-03

1999 8 4,168 -- -- -- 1.00E-03 3.48E-03 1.96E-03

2000 12 3,909 -- -- -- 1.79E-03 4.92E-03 3.07E-03

2001 12 4,023 -- -- -- 1.75E-03 4.79E-03 2.99E-03

2002 10 4,272 -- -- -- 1.31E-03 3.96E-03 2.37E-03

2003 17 4,329 -- -- -- 2.50E-03 5.80E-03 3.90E-03

2004 14 4,338 -- -- -- 1.97E-03 4.99E-03 3.22E-03

2005 16 4,383 -- -- -- 2.30E-03 5.47E-03 3.63E-03

2006 9 4,371 -- -- -- 1.12E-03 3.60E-03 2.09E-03

2007 11 4,288 -- -- -- 1.47E-03 4.23E-03 2.58E-03

2008 8 4,331 -- -- -- 9.65E-04 3.35E-03 1.89E-03

2009 15 4,224 3.79E-03 2.53E-03 5.68E-03 2.20E-03 5.40E-03 3.53E-03

2010 15 4,091 3.57E-03 2.54E-03 5.02E-03 2.27E-03 5.57E-03 3.64E-03

2011 19 4,127 3.37E-03 2.53E-03 4.49E-03 3.00E-03 6.63E-03 4.55E-03

2012 14 4,035 3.18E-03 2.48E-03 4.07E-03 2.11E-03 5.35E-03 3.45E-03

2013 6 4,103 2.99E-03 2.37E-03 3.78E-03 6.91E-04 2.93E-03 1.52E-03

2014 12 4,081 2.82E-03 2.20E-03 3.62E-03 1.72E-03 4.72E-03 2.95E-03

2015 13 4,022 2.66E-03 2.00E-03 3.54E-03 1.93E-03 5.08E-03 3.23E-03

2016 13 4,039 2.51E-03 1.79E-03 3.52E-03 1.92E-03 5.06E-03 3.21E-03

2017 10 3,987 2.36E-03 1.58E-03 3.54E-03 1.40E-03 4.23E-03 2.53E-03

2018 7 3,957 2.23E-03 1.39E-03 3.57E-03 8.81E-04 3.35E-03 1.82E-03

Total 259 87,362



Table 17. Plot data for Figure 2, failure probability estimate trend for EPS EDG FTLR



Mean Lower (5%)

Upper (95%)

Lower (5%)

Upper (95%) Mean

2015 Update 1.18E-03 7.42E-03 3.73E-03

1998 14 3,688 -- -- -- 2.32E-03 5.88E-03 3.79E-03

1999 5 3,648 -- -- -- 6.05E-04 2.95E-03 1.45E-03

2000 8 3,585 -- -- -- 1.17E-03 4.05E-03 2.29E-03

2001 13 3,715 -- -- -- 2.10E-03 5.52E-03 3.51E-03

2002 14 3,638 -- -- -- 2.35E-03 5.96E-03 3.84E-03

2003 15 3,790 -- -- -- 2.46E-03 6.03E-03 3.95E-03

2004 10 3,822 -- -- -- 1.47E-03 4.44E-03 2.65E-03

2005 15 3,784 -- -- -- 2.46E-03 6.04E-03 3.96E-03

2006 15 3,757 -- -- -- 2.48E-03 6.08E-03 3.98E-03

2007 21 3,645 -- -- -- 3.83E-03 8.15E-03 5.69E-03

2008 16 3,729 -- -- -- 2.70E-03 6.44E-03 4.27E-03

2009 18 3,668 4.30E-03 3.35E-03 5.51E-03 3.17E-03 7.17E-03 4.87E-03

2010 11 3,571 4.14E-03 3.36E-03 5.11E-03 1.77E-03 5.08E-03 3.10E-03

2011 16 3,605 3.99E-03 3.35E-03 4.76E-03 2.79E-03 6.65E-03 4.41E-03

2012 17 3,569 3.84E-03 3.30E-03 4.47E-03 3.04E-03 7.04E-03 4.73E-03

2013 12 3,577 3.70E-03 3.22E-03 4.26E-03 1.97E-03 5.40E-03 3.37E-03

2014 12 3,603 3.57E-03 3.08E-03 4.13E-03 1.96E-03 5.36E-03 3.34E-03

2015 13 3,534 3.44E-03 2.91E-03 4.06E-03 2.20E-03 5.79E-03 3.68E-03

2016 11 3,527 3.31E-03 2.72E-03 4.03E-03 1.79E-03 5.14E-03 3.14E-03

2017 9 3,484 3.19E-03 2.53E-03 4.03E-03 1.40E-03 4.51E-03 2.63E-03

2018 13 3,484 3.07E-03 2.34E-03 4.04E-03 2.23E-03 5.87E-03 3.73E-03

Total 278 76,422



Table 18. Plot data for Figure 3, failure rate estimate trend for EPS EDG FTR>1H

Year Failures Run Time

(hr)


Mean Lower (5%)

Upper (95%)

Lower (5%)

Upper (95%) Mean

2015 Update 3.78E-04 3.34E-03 1.54E-03

1998 4 10,000 -- -- -- 1.61E-04 9.55E-04 4.37E-04

1999 1 10,319 -- -- -- 1.66E-05 5.21E-04 1.41E-04

2000 7 11,538 -- -- -- 3.07E-04 1.16E-03 6.33E-04

2001 2 11,977 -- -- -- 4.66E-05 5.73E-04 2.04E-04

2002 7 12,341 -- -- -- 2.87E-04 1.09E-03 5.93E-04

2003 10 11,699 -- -- -- 4.83E-04 1.47E-03 8.75E-04

2004 13 11,747 -- -- -- 6.70E-04 1.77E-03 1.12E-03

2005 14 12,307 -- -- -- 7.02E-04 1.78E-03 1.15E-03

2006 4 11,369 -- -- -- 1.42E-04 8.43E-04 3.86E-04

2007 17 11,365 -- -- -- 9.63E-04 2.24E-03 1.50E-03

2008 20 11,238 -- -- -- 1.18E-03 2.57E-03 1.78E-03

2009 8 11,223 1.21E-03 7.09E-04 2.06E-03 3.76E-04 1.31E-03 7.37E-04

2010 13 10,777 1.28E-03 8.14E-04 2.02E-03 7.29E-04 1.92E-03 1.22E-03

2011 21 11,810 1.36E-03 9.26E-04 2.00E-03 1.20E-03 2.54E-03 1.77E-03

2012 11 7,257 1.44E-03 1.04E-03 2.00E-03 8.66E-04 2.49E-03 1.52E-03

2013 17 7,796 1.53E-03 1.14E-03 2.06E-03 1.39E-03 3.22E-03 2.16E-03

2014 17 7,133 1.62E-03 1.21E-03 2.18E-03 1.51E-03 3.51E-03 2.35E-03

2015 12 7,614 1.72E-03 1.24E-03 2.38E-03 9.23E-04 2.53E-03 1.58E-03

2016 10 7,425 1.82E-03 1.25E-03 2.66E-03 7.50E-04 2.28E-03 1.36E-03

2017 22 7,223 1.93E-03 1.24E-03 3.02E-03 2.03E-03 4.25E-03 2.99E-03

2018 10 7,670 2.05E-03 1.21E-03 3.46E-03 7.27E-04 2.21E-03 1.32E-03

Total 240 211,828



Table 19. Plot data for Figure 4, failure probability estimate trend for HPCS EDG FTS



Mean Lower (5%)

Upper (95%)

Lower (5%)

Upper (95%) Mean

2015 Update 2.19E-04 2.12E-03 9.56E-04

1998 0 151 -- -- -- 4.51E-06 8.94E-03 1.15E-03

1999 1 187 -- -- -- 3.73E-04 1.17E-02 3.18E-03

2000 0 145 -- -- -- 4.58E-06 9.06E-03 1.16E-03

2001 0 154 -- -- -- 4.48E-06 8.88E-03 1.14E-03

2002 0 140 -- -- -- 4.62E-06 9.16E-03 1.18E-03

2003 0 152 -- -- -- 4.50E-06 8.91E-03 1.14E-03

2004 0 139 -- -- -- 4.63E-06 9.17E-03 1.18E-03

2005 0 133 -- -- -- 4.70E-06 9.31E-03 1.20E-03

2006 0 136 -- -- -- 4.66E-06 9.24E-03 1.19E-03

2007 0 126 -- -- -- 4.79E-06 9.48E-03 1.22E-03

2008 1 152 -- -- -- 4.03E-04 1.26E-02 3.44E-03

2009 0 138 1.08E-03 5.87E-04 2.00E-03 4.65E-06 9.20E-03 1.18E-03

2010 0 151 1.16E-03 6.89E-04 1.95E-03 4.51E-06 8.93E-03 1.15E-03

2011 0 152 1.24E-03 8.00E-04 1.92E-03 4.50E-06 8.91E-03 1.14E-03

2012 0 128 1.32E-03 9.10E-04 1.92E-03 4.76E-06 9.43E-03 1.21E-03

2013 0 158 1.41E-03 1.00E-03 1.99E-03 4.44E-06 8.80E-03 1.13E-03

2014 1 136 1.51E-03 1.06E-03 2.15E-03 4.19E-04 1.31E-02 3.57E-03

2015 0 140 1.61E-03 1.08E-03 2.41E-03 4.63E-06 9.17E-03 1.18E-03

2016 0 138 1.73E-03 1.08E-03 2.77E-03 4.65E-06 9.21E-03 1.18E-03

2017 0 139 1.84E-03 1.05E-03 3.23E-03 4.64E-06 9.19E-03 1.18E-03

2018 1 144 1.97E-03 1.02E-03 3.80E-03 4.11E-04 1.29E-02 3.50E-03

Total 4 3,038



Table 20. Plot data for Figure 5, failure probability estimate trend for HPCS EDG FTLR



Mean Lower (5%)

Upper (95%)

Lower (5%)

Upper (95%) Mean

2015 Update 4.25E-04 2.16E-03 1.15E-03

1998 0 106 -- -- -- 8.89E-06 1.76E-02 2.26E-03

1999 1 131 -- -- -- 7.16E-04 2.23E-02 6.10E-03

2000 0 121 -- -- -- 8.33E-06 1.64E-02 2.12E-03

2001 0 125 -- -- -- 8.17E-06 1.61E-02 2.08E-03

2002 1 124 -- -- -- 7.35E-04 2.29E-02 6.27E-03

2003 0 129 -- -- -- 8.06E-06 1.59E-02 2.05E-03

2004 1 129 -- -- -- 7.20E-04 2.25E-02 6.14E-03

2005 0 119 -- -- -- 8.38E-06 1.65E-02 2.13E-03

2006 0 124 -- -- -- 8.20E-06 1.62E-02 2.09E-03

2007 0 117 -- -- -- 8.45E-06 1.67E-02 2.15E-03

2008 0 139 -- -- -- 7.72E-06 1.53E-02 1.96E-03

2009 0 118 3.28E-03 1.32E-03 8.12E-03 8.42E-06 1.66E-02 2.14E-03

2010 1 132 3.34E-03 1.55E-03 7.22E-03 7.12E-04 2.22E-02 6.07E-03

2011 0 137 3.41E-03 1.78E-03 6.52E-03 7.79E-06 1.54E-02 1.98E-03

2012 1 117 3.48E-03 2.00E-03 6.04E-03 7.58E-04 2.36E-02 6.46E-03

2013 0 137 3.55E-03 2.16E-03 5.83E-03 7.79E-06 1.54E-02 1.98E-03

2014 1 118 3.62E-03 2.20E-03 5.96E-03 7.55E-04 2.35E-02 6.43E-03

2015 1 127 3.70E-03 2.12E-03 6.42E-03 7.27E-04 2.27E-02 6.19E-03

2016 0 126 3.77E-03 1.97E-03 7.22E-03 8.14E-06 1.61E-02 2.07E-03

2017 0 125 3.85E-03 1.77E-03 8.32E-03 8.18E-06 1.62E-02 2.08E-03

2018 1 129 3.93E-03 1.58E-03 9.74E-03 7.21E-04 2.25E-02 6.14E-03

Total 8 2,630



Table 21. Plot data for Figure 6, failure rate estimate trend for HPCS EDG FTR>1H

Year Failures Run Time

(hr)


Mean Lower (5%)

Upper (95%)

Lower (5%)

Upper (95%) Mean

2015 Update 4.25E-04 2.16E-03 1.15E-03

1998 0 296 -- -- -- 3.65E-06 7.26E-03 9.29E-04

1999 1 459 -- -- -- 2.51E-04 7.89E-03 2.14E-03

2000 0 348 -- -- -- 3.33E-06 6.62E-03 8.47E-04

2001 0 361 -- -- -- 3.26E-06 6.48E-03 8.29E-04

2002 0 327 -- -- -- 3.45E-06 6.86E-03 8.77E-04

2003 0 391 -- -- -- 3.11E-06 6.17E-03 7.90E-04

2004 0 331 -- -- -- 3.43E-06 6.81E-03 8.71E-04

2005 1 376 -- -- -- 2.84E-04 8.94E-03 2.42E-03

2006 0 379 -- -- -- 3.16E-06 6.29E-03 8.05E-04

2007 0 307 -- -- -- 3.58E-06 7.11E-03 9.10E-04

2008 0 452 -- -- -- 2.83E-06 5.62E-03 7.20E-04

2009 0 381 9.69E-04 3.93E-04 2.39E-03 3.15E-06 6.27E-03 8.02E-04

2010 1 392 1.09E-03 5.06E-04 2.34E-03 2.77E-04 8.72E-03 2.36E-03

2011 0 425 1.22E-03 6.41E-04 2.32E-03 2.95E-06 5.86E-03 7.50E-04

2012 0 162 1.37E-03 7.92E-04 2.36E-03 4.87E-06 9.67E-03 1.24E-03

2013 1 298 1.53E-03 9.38E-04 2.51E-03 3.25E-04 1.02E-02 2.78E-03

2014 0 194 1.72E-03 1.05E-03 2.82E-03 4.51E-06 8.96E-03 1.15E-03

2015 0 228 1.93E-03 1.11E-03 3.35E-03 4.18E-06 8.30E-03 1.06E-03

2016 1 205 2.17E-03 1.13E-03 4.15E-03 3.93E-04 1.24E-02 3.35E-03

2017 0 195 2.43E-03 1.12E-03 5.27E-03 4.50E-06 8.94E-03 1.14E-03

2018 2 186 2.73E-03 1.09E-03 6.79E-03 1.34E-03 1.64E-02 5.83E-03

Total 7 6,694



Table 22. Plot data for Figure 7, EPS EDG UA trend

Year UA Hours Critical Hours


Mean Lower (5%)

Upper (95%)

Lower (5%)

Upper (95%) Mean

2015 Update 4.44E-03 2.52E-02 1.48E-02

1998 21,235 1,874,166 -- -- -- 2.01E-03 2.42E-02 1.04E-02

1999 22,769 2,005,223 -- -- -- 2.55E-03 2.42E-02 1.10E-02

2000 18,409 2,042,467 -- -- -- 2.27E-03 2.08E-02 9.53E-03

2001 19,233 2,075,373 -- -- -- 1.54E-03 2.36E-02 9.72E-03

2002 24,631 2,093,196 -- -- -- 2.31E-03 2.78E-02 1.20E-02

2003 28,961 2,047,203 -- -- -- 1.91E-03 3.49E-02 1.39E-02

2004 29,617 2,099,392 -- -- -- 1.24E-03 3.71E-02 1.36E-02

2005 26,350 2,070,016 -- -- -- 2.81E-03 2.90E-02 1.29E-02

2006 28,713 2,083,212 -- -- -- 1.73E-03 3.36E-02 1.33E-02

2007 34,106 2,104,023 -- -- -- 2.41E-03 3.99E-02 1.62E-02

2008 31,755 2,089,978 -- -- -- 2.87E-03 3.53E-02 1.52E-02

2009 33,204 2,059,429 1.60E-02 1.45E-02 1.76E-02 2.86E-03 3.79E-02 1.61E-02

2010 30,037 2,081,690 1.58E-02 1.46E-02 1.70E-02 3.32E-03 3.16E-02 1.44E-02

2011 36,401 2,023,478 1.56E-02 1.47E-02 1.64E-02 2.95E-03 4.18E-02 1.75E-02

2012 32,470 1,977,596 1.53E-02 1.48E-02 1.58E-02 3.02E-03 3.68E-02 1.59E-02

2013 30,642 2,007,371 1.51E-02 1.49E-02 1.53E-02 2.29E-03 3.34E-02 1.39E-02

2014 28,292 2,027,147 1.49E-02 1.47E-02 1.50E-02 3.15E-03 3.19E-02 1.43E-02

2015 30,706 2,008,809 1.46E-02 1.42E-02 1.51E-02 3.65E-03 3.36E-02 1.54E-02

2016 29,859 2,025,233 1.44E-02 1.37E-02 1.52E-02 2.95E-04 5.06E-02 1.51E-02

2017 30,463 1,997,343 1.42E-02 1.32E-02 1.53E-02 2.00E-03 3.69E-02 1.47E-02

2018 25,907 1,990,438 1.40E-02 1.27E-02 1.54E-02 2.88E-03 2.93E-02 1.31E-02

Total 593,761 42,782,783



Table 23. Plot data for Figure 8, HPCS EDG UA trend

Year UA Hours Critical Hours


Mean Lower (5%)

Upper (95%)

Lower (5%)

Upper (95%) Mean

2015 Update 7.59E-03 1.58E-02 1.17E-02

1998 255 42,029 -- -- -- 4.24E-04 1.15E-02 4.27E-03

1999 760 55,565 -- -- -- 7.09E-04 3.99E-02 1.35E-02

2000 959 65,705 -- -- -- 8.40E-04 4.33E-02 1.48E-02

2001 474 65,093 -- -- -- 1.27E-03 1.69E-02 7.13E-03

2002 431 65,329 -- -- -- 1.14E-03 1.59E-02 6.66E-03

2003 825 65,040 -- -- -- 6.07E-03 2.11E-02 1.26E-02

2004 855 65,589 -- -- -- 4.00E-03 2.63E-02 1.31E-02

2005 610 64,383 -- -- -- 3.59E-03 1.75E-02 9.42E-03

2006 453 66,949 -- -- -- 1.85E-03 1.40E-02 6.71E-03

2007 592 64,512 -- -- -- 1.92E-03 2.07E-02 9.14E-03

2008 861 65,262 -- -- -- 2.31E-03 3.24E-02 1.36E-02

2009 519 63,966 1.16E-02 7.37E-03 1.58E-02 2.54E-03 1.63E-02 8.13E-03

2010 1,050 67,158 1.21E-02 8.83E-03 1.54E-02 2.34E-03 3.80E-02 1.55E-02

2011 991 62,329 1.26E-02 1.03E-02 1.50E-02 4.37E-03 3.29E-02 1.58E-02

2012 815 64,557 1.32E-02 1.17E-02 1.46E-02 2.20E-03 2.92E-02 1.24E-02

2013 952 64,142 1.37E-02 1.32E-02 1.41E-02 2.33E-03 3.51E-02 1.45E-02

2014 1,012 66,677 1.42E-02 1.37E-02 1.47E-02 2.39E-03 3.66E-02 1.51E-02

2015 627 65,277 1.47E-02 1.33E-02 1.61E-02 1.40E-03 2.37E-02 9.56E-03

2016 884 62,704 1.52E-02 1.29E-02 1.76E-02 3.09E-03 3.03E-02 1.37E-02

2017 827 63,353 1.58E-02 1.25E-02 1.90E-02 1.76E-03 3.45E-02 1.36E-02

2018 1,346 64,262 1.63E-02 1.21E-02 2.05E-02 1.46E-03 6.04E-02 2.12E-02

Total 16,097 1,329,882



Table 24. Plot data for Figure 9, EPS EDG unreliability trend (8-hour mission)

Year


Mean Lower (5%) Upper (95%) Lower (5%) Upper (95%) Mean

1998 -- -- -- 1.22E-02 3.54E-02 2.11E-02

1999 -- -- -- 6.16E-03 2.93E-02 1.49E-02

2000 -- -- -- 1.16E-02 3.45E-02 2.08E-02

2001 -- -- -- 9.85E-03 3.18E-02 1.87E-02

2002 -- -- -- 1.15E-02 3.04E-02 1.94E-02

2003 -- -- -- 1.48E-02 3.62E-02 2.37E-02

2004 -- -- -- 1.41E-02 3.72E-02 2.30E-02

2005 -- -- -- 1.56E-02 3.74E-02 2.45E-02

2006 -- -- -- 1.10E-02 3.68E-02 2.09E-02

2007 -- -- -- 1.93E-02 4.50E-02 2.99E-02

2008 -- -- -- 1.87E-02 5.30E-02 3.17E-02

2009 2.97E-02 2.55E-02 3.47E-02 1.39E-02 4.99E-02 2.71E-02

2010 3.02E-02 2.65E-02 3.45E-02 1.51E-02 5.34E-02 2.88E-02

2011 3.07E-02 2.75E-02 3.43E-02 2.05E-02 5.99E-02 3.46E-02

2012 3.12E-02 2.84E-02 3.43E-02 1.95E-02 4.93E-02 3.15E-02

2013 3.18E-02 2.92E-02 3.46E-02 1.95E-02 4.86E-02 3.20E-02

2014 3.23E-02 2.97E-02 3.51E-02 2.16E-02 5.56E-02 3.55E-02

2015 3.28E-02 2.99E-02 3.61E-02 1.73E-02 5.16E-02 3.09E-02

2016 3.34E-02 2.99E-02 3.73E-02 1.65E-02 5.33E-02 3.11E-02

2017 3.39E-02 2.97E-02 3.87E-02 2.53E-02 6.66E-02 4.21E-02

2018 3.45E-02 2.95E-02 4.03E-02 1.57E-02 4.84E-02 2.89E-02



Table 25. Plot data for Figure 10, HPCS EDG unreliability trend (8-hour mission)

Year


Mean Lower (5%) Upper (95%) Lower (5%) Upper (95%) Mean

1998 -- -- -- 2.88E-03 2.79E-02 1.20E-02

1999 -- -- -- 9.87E-03 4.38E-02 2.42E-02

2000 -- -- -- 4.03E-03 5.25E-02 2.15E-02

2001 -- -- -- 3.79E-03 4.87E-02 2.09E-02

2002 -- -- -- 6.50E-03 6.06E-02 2.62E-02

2003 -- -- -- 4.53E-03 5.36E-02 2.27E-02

2004 -- -- -- 6.34E-03 3.50E-02 1.81E-02

2005 -- -- -- 8.45E-03 4.28E-02 2.25E-02

2006 -- -- -- 3.88E-03 2.98E-02 1.38E-02

2007 -- -- -- 4.04E-03 3.01E-02 1.41E-02

2008 -- -- -- 1.06E-02 3.55E-02 2.11E-02

2009 2.07E-02 1.38E-02 3.11E-02 9.47E-03 3.36E-02 1.95E-02

2010 2.14E-02 1.51E-02 3.02E-02 1.34E-02 5.41E-02 3.14E-02

2011 2.21E-02 1.65E-02 2.95E-02 7.67E-03 3.70E-02 1.95E-02

2012 2.28E-02 1.78E-02 2.91E-02 9.17E-03 4.20E-02 2.19E-02

2013 2.35E-02 1.88E-02 2.93E-02 1.05E-02 4.78E-02 2.56E-02

2014 2.42E-02 1.94E-02 3.03E-02 8.95E-03 3.79E-02 2.06E-02

2015 2.50E-02 1.95E-02 3.20E-02 6.56E-03 3.62E-02 1.81E-02

2016 2.58E-02 1.93E-02 3.45E-02 9.80E-03 5.79E-02 2.86E-02

2017 2.66E-02 1.89E-02 3.76E-02 5.29E-03 3.62E-02 1.73E-02

2018 2.75E-02 1.83E-02 4.13E-02 2.12E-02 8.53E-02 4.87E-02



Table 26. Plot data for Figure 11, frequency of start demands (demands per reactor year) trend for EPS and

HPCS EDGs

Year Demands Reactor Years


Mean Lower (5%)

Upper (95%)

Lower (5%)

Upper (95%) Mean

1998 4,436 95.0 -- -- -- 4.55E+01 4.79E+01 4.67E+01

1999 4,355 95.0 -- -- -- 4.47E+01 4.70E+01 4.58E+01

2000 4,054 95.3 -- -- -- 4.15E+01 4.37E+01 4.26E+01

2001 4,177 95.0 -- -- -- 4.29E+01 4.51E+01 4.40E+01

2002 4,412 95.0 -- -- -- 4.53E+01 4.76E+01 4.64E+01

2003 4,481 95.0 -- -- -- 4.60E+01 4.83E+01 4.72E+01

2004 4,477 95.3 -- -- -- 4.58E+01 4.82E+01 4.70E+01

2005 4,516 95.0 -- -- -- 4.64E+01 4.87E+01 4.75E+01

2006 4,508 95.0 -- -- -- 4.63E+01 4.86E+01 4.75E+01

2007 4,413 95.0 -- -- -- 4.53E+01 4.76E+01 4.65E+01

2008 4,483 95.3 -- -- -- 4.59E+01 4.82E+01 4.71E+01

2009 4,362 95.0 4.49E+01 4.37E+01 4.61E+01 4.48E+01 4.71E+01 4.59E+01

2010 4,242 95.0 4.51E+01 4.41E+01 4.61E+01 4.35E+01 4.58E+01 4.47E+01

2011 4,279 95.0 4.52E+01 4.44E+01 4.61E+01 4.39E+01 4.62E+01 4.50E+01

2012 4,162 95.3 4.54E+01 4.47E+01 4.61E+01 4.26E+01 4.48E+01 4.37E+01

2013 4,261 92.6 4.56E+01 4.49E+01 4.62E+01 4.49E+01 4.72E+01 4.60E+01

2014 4,217 91.0 4.58E+01 4.51E+01 4.64E+01 4.52E+01 4.75E+01 4.63E+01

2015 4,162 90.0 4.59E+01 4.52E+01 4.67E+01 4.51E+01 4.74E+01 4.62E+01

2016 4,177 90.0 4.61E+01 4.53E+01 4.70E+01 4.52E+01 4.76E+01 4.64E+01

2017 4,126 89.0 4.63E+01 4.53E+01 4.73E+01 4.52E+01 4.76E+01 4.64E+01

2018 4,100 88.7 4.65E+01 4.53E+01 4.77E+01 4.50E+01 4.74E+01 4.62E+01

Total 90,400 1,967.4



Table 27. Plot data for Figure 12, frequency of FTLR demands (demands per reactor year) trend for EPS

and HPCS EDGs

Year Demands Reactor Years


Mean Lower (5%)

Upper (95%)

Lower (5%)

Upper (95%) Mean

1998 3,793 95.0 -- -- -- 3.89E+01 4.10E+01 3.99E+01

1999 3,778 95.0 -- -- -- 3.87E+01 4.09E+01 3.98E+01

2000 3,706 95.3 -- -- -- 3.79E+01 4.00E+01 3.89E+01

2001 3,840 95.0 -- -- -- 3.94E+01 4.15E+01 4.04E+01

2002 3,762 95.0 -- -- -- 3.85E+01 4.07E+01 3.96E+01

2003 3,919 95.0 -- -- -- 4.02E+01 4.24E+01 4.13E+01

2004 3,951 95.3 -- -- -- 4.04E+01 4.26E+01 4.15E+01

2005 3,903 95.0 -- -- -- 4.00E+01 4.22E+01 4.11E+01

2006 3,882 95.0 -- -- -- 3.98E+01 4.20E+01 4.09E+01

2007 3,762 95.0 -- -- -- 3.85E+01 4.07E+01 3.96E+01

2008 3,868 95.3 -- -- -- 3.95E+01 4.17E+01 4.06E+01

2009 3,786 95.0 3.92E+01 3.84E+01 4.00E+01 3.88E+01 4.09E+01 3.99E+01

2010 3,704 95.0 3.94E+01 3.87E+01 4.01E+01 3.79E+01 4.01E+01 3.90E+01

2011 3,742 95.0 3.96E+01 3.90E+01 4.02E+01 3.83E+01 4.05E+01 3.94E+01

2012 3,686 95.3 3.97E+01 3.92E+01 4.03E+01 3.76E+01 3.98E+01 3.87E+01

2013 3,714 92.6 3.99E+01 3.95E+01 4.04E+01 3.90E+01 4.12E+01 4.01E+01

2014 3,721 91.0 4.01E+01 3.97E+01 4.06E+01 3.98E+01 4.20E+01 4.09E+01

2015 3,661 90.0 4.03E+01 3.98E+01 4.08E+01 3.96E+01 4.18E+01 4.07E+01

2016 3,653 90.0 4.05E+01 3.99E+01 4.11E+01 3.95E+01 4.17E+01 4.06E+01

2017 3,609 89.0 4.07E+01 4.00E+01 4.14E+01 3.94E+01 4.17E+01 4.05E+01

2018 3,613 88.7 4.09E+01 4.00E+01 4.18E+01 3.96E+01 4.18E+01 4.07E+01

Total 79,052 1,967.4



Table 28. Plot data for Figure 13, frequency of run > 1H hours (hours per reactor year) trend for EPS and

HPCS EDG

Year Run

Hours Reactor Years


Mean Lower (5%)

Upper (95%)

Lower (5%)

Upper (95%) Mean

1998 10,296 95.0 -- -- -- 1.07E+02 1.10E+02 1.08E+02

1999 10,779 95.0 -- -- -- 1.12E+02 1.15E+02 1.13E+02

2000 11,886 95.3 -- -- -- 1.23E+02 1.27E+02 1.25E+02

2001 12,338 95.0 -- -- -- 1.28E+02 1.32E+02 1.30E+02

2002 12,668 95.0 -- -- -- 1.31E+02 1.35E+02 1.33E+02

2003 12,090 95.0 -- -- -- 1.25E+02 1.29E+02 1.27E+02

2004 12,078 95.3 -- -- -- 1.25E+02 1.29E+02 1.27E+02

2005 12,683 95.0 -- -- -- 1.32E+02 1.35E+02 1.34E+02

2006 11,748 95.0 -- -- -- 1.22E+02 1.26E+02 1.24E+02

2007 11,672 95.0 -- -- -- 1.21E+02 1.25E+02 1.23E+02

2008 11,690 95.3 -- -- -- 1.21E+02 1.25E+02 1.23E+02

2009 11,604 95.0 1.16E+02 9.53E+01 1.42E+02 1.20E+02 1.24E+02 1.22E+02

2010 11,169 95.0 1.11E+02 9.38E+01 1.31E+02 1.16E+02 1.19E+02 1.18E+02

2011 12,234 95.0 1.06E+02 9.19E+01 1.22E+02 1.27E+02 1.31E+02 1.29E+02

2012 7,419 95.3 1.01E+02 8.94E+01 1.14E+02 7.64E+01 7.94E+01 7.79E+01

2013 8,094 92.6 9.64E+01 8.59E+01 1.08E+02 8.58E+01 8.91E+01 8.74E+01

2014 7,327 91.0 9.19E+01 8.14E+01 1.04E+02 7.90E+01 8.21E+01 8.05E+01

2015 7,843 90.0 8.77E+01 7.63E+01 1.01E+02 8.55E+01 8.88E+01 8.71E+01

2016 7,630 90.0 8.37E+01 7.09E+01 9.88E+01 8.32E+01 8.64E+01 8.48E+01

2017 7,418 89.0 7.98E+01 6.55E+01 9.72E+01 8.18E+01 8.50E+01 8.33E+01

2018 7,856 88.7 7.62E+01 6.04E+01 9.60E+01 8.69E+01 9.02E+01 8.85E+01

Total 218,522 1,967.4



Table 29. Plot data for Figure 14, frequency of FTS events (events per reactor year) trend for EPS and

HPCS EDGs

Year Failures Reactor Years


Mean Lower (5%)

Upper (95%)

Lower (5%)

Upper (95%) Mean

1998 18 95.0 -- -- -- 1.22E-01 2.77E-01 1.88E-01

1999 9 95.0 -- -- -- 5.13E-02 1.66E-01 9.63E-02

2000 12 95.3 -- -- -- 7.39E-02 2.03E-01 1.26E-01

2001 12 95.0 -- -- -- 7.41E-02 2.03E-01 1.27E-01

2002 10 95.0 -- -- -- 5.88E-02 1.78E-01 1.06E-01

2003 17 95.0 -- -- -- 1.14E-01 2.65E-01 1.77E-01

2004 14 95.3 -- -- -- 8.95E-02 2.27E-01 1.47E-01

2005 16 95.0 -- -- -- 1.06E-01 2.52E-01 1.67E-01

2006 9 95.0 -- -- -- 5.13E-02 1.66E-01 9.63E-02

2007 11 95.0 -- -- -- 6.64E-02 1.91E-01 1.17E-01

2008 9 95.3 -- -- -- 5.11E-02 1.65E-01 9.61E-02

2009 15 95.0 1.63E-01 1.10E-01 2.43E-01 9.77E-02 2.40E-01 1.57E-01

2010 15 95.0 1.56E-01 1.11E-01 2.17E-01 9.77E-02 2.40E-01 1.57E-01

2011 19 95.0 1.48E-01 1.12E-01 1.96E-01 1.30E-01 2.89E-01 1.98E-01

2012 14 95.3 1.41E-01 1.11E-01 1.80E-01 8.95E-02 2.27E-01 1.47E-01

2013 6 92.6 1.34E-01 1.07E-01 1.69E-01 3.06E-02 1.30E-01 6.76E-02

2014 13 91.0 1.28E-01 1.01E-01 1.63E-01 8.53E-02 2.25E-01 1.43E-01

2015 13 90.0 1.22E-01 9.24E-02 1.60E-01 8.62E-02 2.27E-01 1.44E-01

2016 13 90.0 1.16E-01 8.35E-02 1.61E-01 8.62E-02 2.27E-01 1.44E-01

2017 10 89.0 1.10E-01 7.48E-02 1.63E-01 6.26E-02 1.90E-01 1.13E-01

2018 8 88.7 1.05E-01 6.66E-02 1.66E-01 4.69E-02 1.63E-01 9.20E-02

Total 263 1,967.4



Table 30. Plot data for Figure 15, frequency of FTLR events (events per reactor year) trend for EPS and

HPCS EDGs



Mean Lower (5%)

Upper (95%)

Lower (5%)

Upper (95%) Mean

1998 14 95.0 -- -- -- 8.99E-02 2.28E-01 1.47E-01

1999 5 95.0 -- -- -- 2.32E-02 1.14E-01 5.59E-02

2000 8 95.3 -- -- -- 4.39E-02 1.53E-01 8.61E-02

2001 13 95.0 -- -- -- 8.20E-02 2.16E-01 1.37E-01

2002 14 95.0 -- -- -- 8.99E-02 2.28E-01 1.47E-01

2003 15 95.0 -- -- -- 9.79E-02 2.41E-01 1.57E-01

2004 10 95.3 -- -- -- 5.87E-02 1.78E-01 1.06E-01

2005 15 95.0 -- -- -- 9.79E-02 2.41E-01 1.57E-01

2006 15 95.0 -- -- -- 9.79E-02 2.41E-01 1.57E-01

2007 21 95.0 -- -- -- 1.47E-01 3.13E-01 2.18E-01

2008 16 95.3 -- -- -- 1.06E-01 2.52E-01 1.67E-01

2009 18 95.0 1.63E-01 1.27E-01 2.09E-01 1.22E-01 2.77E-01 1.88E-01

2010 11 95.0 1.58E-01 1.28E-01 1.94E-01 6.65E-02 1.91E-01 1.17E-01

2011 16 95.0 1.52E-01 1.28E-01 1.82E-01 1.06E-01 2.53E-01 1.68E-01

2012 17 95.3 1.48E-01 1.27E-01 1.72E-01 1.14E-01 2.64E-01 1.77E-01

2013 12 92.6 1.43E-01 1.24E-01 1.64E-01 7.61E-02 2.09E-01 1.30E-01

2014 12 91.0 1.38E-01 1.19E-01 1.60E-01 7.73E-02 2.12E-01 1.32E-01

2015 13 90.0 1.34E-01 1.13E-01 1.58E-01 8.64E-02 2.28E-01 1.44E-01

2016 11 90.0 1.30E-01 1.06E-01 1.58E-01 7.00E-02 2.01E-01 1.23E-01

2017 9 89.0 1.25E-01 9.92E-02 1.59E-01 5.47E-02 1.77E-01 1.03E-01

2018 13 88.7 1.21E-01 9.21E-02 1.60E-01 8.76E-02 2.31E-01 1.46E-01

Total 278 1,967.4



Table 31. Plot data for Figure 16, frequency of FTR>1H events (events per reactor year) trend for EPS and

HPCS EDGs



Mean Lower (5%)

Upper (95%)

Lower (5%)

Upper (95%) Mean

1998 4 95.0 -- -- -- 1.69E-02 1.00E-01 4.58E-02

1999 1 95.0 -- -- -- 1.79E-03 5.63E-02 1.53E-02

2000 7 95.3 -- -- -- 3.69E-02 1.40E-01 7.61E-02

2001 2 95.0 -- -- -- 5.83E-03 7.16E-02 2.54E-02

2002 7 95.0 -- -- -- 3.70E-02 1.40E-01 7.63E-02

2003 10 95.0 -- -- -- 5.90E-02 1.79E-01 1.07E-01

2004 13 95.3 -- -- -- 8.20E-02 2.16E-01 1.37E-01

2005 14 95.0 -- -- -- 9.01E-02 2.29E-01 1.48E-01

2006 4 95.0 -- -- -- 1.69E-02 1.00E-01 4.58E-02

2007 17 95.0 -- -- -- 1.14E-01 2.66E-01 1.78E-01

2008 20 95.3 -- -- -- 1.39E-01 3.01E-01 2.08E-01

2009 8 95.0 1.33E-01 7.95E-02 2.23E-01 4.41E-02 1.53E-01 8.65E-02

2010 13 95.0 1.36E-01 8.77E-02 2.10E-01 8.22E-02 2.17E-01 1.37E-01

2011 21 95.0 1.39E-01 9.59E-02 2.00E-01 1.47E-01 3.14E-01 2.19E-01

2012 11 95.3 1.41E-01 1.03E-01 1.93E-01 6.64E-02 1.91E-01 1.17E-01

2013 17 92.6 1.44E-01 1.09E-01 1.91E-01 1.17E-01 2.72E-01 1.83E-01

2014 17 91.0 1.47E-01 1.12E-01 1.94E-01 1.19E-01 2.77E-01 1.86E-01

2015 12 90.0 1.50E-01 1.11E-01 2.04E-01 7.83E-02 2.15E-01 1.34E-01

2016 10 90.0 1.53E-01 1.07E-01 2.19E-01 6.22E-02 1.89E-01 1.13E-01

2017 22 89.0 1.57E-01 1.02E-01 2.39E-01 1.66E-01 3.47E-01 2.44E-01

2018 10 88.7 1.60E-01 9.68E-02 2.64E-01 6.30E-02 1.91E-01 1.14E-01

Total 240 1,967.4



9. REFERENCES

[1] United States Nuclear Regulatory Commission, " Component Performance Study: Emergency

Diesel Generators 1998-2006," December 2007. [Online]. Available:

https://nrcoe.inl.gov/resultsdb/publicdocs/CompPerf/edg-2006.pdf.

[2] J. A. Schroeder, "Enhanced Component Performance Study: Emergency Diesel Generators 1998-

2016," INL/LTD-17-44204, Idaho National laboratory, April 2018.

[3] C. D. Gentillion, "Overview and Reference Document for Operational Experience Results and

Databases Trending," February 2016. [Online]. Available:

https://nrcoe.inel.gov/resultsdb/publicdocs/Overview-and-Reference.pdf.

[4] J. C. Lane, "NRC Operating Experience (OpE) Programs," Office of Nuclear Regulatory Research,

July 2015. [Online]. Available: http://pbadupws.nrc.gov/docs/ML1518/ML15189A345.pdf.

[Accessed 2015].

[5] Nuclear Energy Institute, "Regulatory Assessment Performance Indicator Guideline," NEI 99-02,

Revision 7, August 2013.

[6] United States Nuclear Regulatory Commission, "Component Reliability Data Sheets Update 2015,"

February 2017. [Online]. Available:

http://nrcoe.inl.gov/resultsdb/publicdocs/AvgPerf/ComponentReliabilityDataSheets2015.pdf.

[7] S. A. Eide, T. E. Wierman, C. D. Gentillon, D. M. Rasmuson and C. L. Atwood, "Industry-Average

Performance for Components and Initiating Events at U.S. Commercial Nuclear Power Plants,"

NUREG/CR-6928, U.S. Nuclear Regulatory Commission, February 2007.

[8] C. L. Atwood, etc., "Handbook of Parameter Estimation for Probabilistic Risk Assessment,"

NUREG/CR-6823, U.S. Nuclear Regulatory Commission, September 2003.

[9] Institute of Nuclear Power Operations, "INPO Consolidated Event System (ICES) Reporting

Requirements and Standards," INPO 12-009, Revision 8, Institute of Nuclear Power Operations,

November 2015.

https://nrcoe.inl.gov/resultsdb/publicdocs/CompPerf/edg-2006.pdf

https://nrcoe.inel.gov/resultsdb/publicdocs/Overview-and-Reference.pdf


enhanced component performance study: emergency diesel

Documents